Abstract
We assessed HbA1c sensitivity and specificity in detecting diabetes using the 75 g oral glucose tolerance test (OGTT) as the gold standard in 1485 subjects aged 22-89 years.
For HbA1c at the recommended ≥6.1% (JDS) cutoff, sensitivity was 56.6% and specificity 98.5%.
Based on receiver operating characteristic (ROC) analysis, the optimal HbA1c cutoff was 6.0% and FPG 106 mg/dl. Combining ≥6.0% HbA1c and ≥106 mg/dl FPG to detect diabetes, sensitivity increased markedly to 94% but specificity dropped to 92%.
Both HbA1c and FPG have low sensitivity of <40% for detecting poor (borderline) glucose tolerance.
The limited sensitivity of HbA1c alone may fail to identify a high proportion ->40%- of those with diabetes, so diabetes screening strategy should combine at least HbA1c and FPG.
